territory score formula

[Aeson]

The more I test the domination limit, the less sense it makes. I finished the GOTM5 earlier today, and went back to see how close I was to the domination limit. Then I counted squares, and played out a lot of scenarios to find the exact tile count needed to trigger domination. Turned out I had guessed pretty close, just 32 tiles off.

In my edited maps that I tested on, coastal tiles didn't seem to count, but on the GOTM map they did. Maybe this has something to do with the fact that I just used coastal and ocean tiles in my edited maps with no sea inbetween? Or maybe the map generation in game is slightly different than in the editor.

I don't want to give exact numbers here, as it would be GOTM spoiler info. I used my Territory # from demographics to figure out how many claimed tiles I had. Anyways, without counting water tiles it was was nowhere near 67%. My Territory was well over the TotalLand tiles.

Territory / TotalLand != 67%

Even when I counted Coastal tiles into the equation it still wasn't anywhere near 67% of Land and Coastal tiles.

Territory / TotalLandCoast != 67%

When I added in Sea tiles it was much closer, but still wasn't 67%.

Territory / TotalLandCoastSea != 67%

The only way that I could get anything near 67% was to take a rough estimate of the number of Sea tiles in my borders, and subtract that from my Territory. Then dividing that by the Total Land, Coast, and Sea tiles came up pretty close. I don't know if Sea tiles even count towards the Territory number in the Demographic screen though, and my estimate was quite rough, was pretty tired of counting tiles.

Territory - ClaimedSea / TotalLandCoastSea = ~67%

I'd put the numbers up on the GOTM5 Spoiler thread, but posting the exact domination limit of the map would be too much of a spoiler.

 

[David In Asia]

do you think maybe this points to a flaw in the game?? surely somebody shouldn't need to be going to all this effort in order to find out something that, in all honesty, should be seen within the game. the domination victory has always been very unclear to me (up until now, thanks to cp) and i'm only on my third game, so i don't pretend to be and expert, but hey, just my small opinion, so be gentle with me

 

[chiefpaco]

Aeson. Good work. I think your formula would still work for my case because I made no sea either. I can try that in my next test. Weird about not including your sea squares though.

David. I agree I would like to see at least an explanation of what we need to control and how much (obviously, otherwise I would never have posted). However, some ppl around here have convinced me that not knowing it exactly is ok too. That won't stop me from trying this, but there's a good argument for either side.

This recent thread started by SirPleb kinda goes through the argument of it without the technical analysis I'm going through here .

http://forums.civfanatics.com/showthread.php?s=&threadid=17210

 

[Badluck]

Aeson, could i see your game??
i would like to test some stuff if you still have a saved files of 5 or so turn before victori. *and the information i need to achieve domination *

i don't play GOTM anyway, and i promise not to give any information to other poeple.

 

[Aeson]

Territory - ClaimedSea / TotalLandCoastSea = ~67%

This should have read: (which is even weirder, only way I could figure out to get 67% as the result)

(Territory - ClaimedSea) / (TotalLandCoast + ClaimedSea) = ~67%

If 67% is the magic number, it must be a formula bug where the number of claimed sea tiles count towards the total number of tiles instead of the players claimed tiles.

In a test I just ran, Sea tiles do count towards the demographic's territory number, but not towards the players claimed tiles when figuring domination. I could claim 31 Land/Coastal tiles, and 1 Sea tile, and domination wouldn't trigger. My Territory count was the same as when 32 Land/Coastal tiles were claimed, which would trigger domination.

---------------------------------------------------------------------------------

Ok, here is the zip file of the .SAV I was using for these tests along with a text file with all the numbers. The Settlers on the larger island are in position to run the 31 Land/Coastal + 1 Sea tile test. The Settler on the small southern island is just there for tests that needed 3 Coastal tiles claimed.

*Download is a CivFanatics GOTM5 SPOILER*

Don't look at this unless you aren't participating or have submitted your GOTM5 already.

 

[Badluck]

how do i destroy a citi??

 

[chiefpaco]

Originally posted by Aeson


In a test I just ran, Sea tiles do count towards the demographic's territory number, but not towards the players claimed tiles when figuring domination. I could claim 31 Land/Coastal tiles, and 1 Sea tile, and domination wouldn't trigger. My Territory count was the same as when 32 Land/Coastal tiles were claimed, which would trigger domination.

I confirmed your sea tile test - I've seen the same thing. As for the other stuff, that sounds whacky. I'll see what I can come up with but using a larger map.

 

[David In Asia]

will go over that thread now for my twopence worth......sorry to butt in on your discussion!

 

[SirPleb]

Originally posted by Aeson
(Territory - ClaimedSea) / (TotalLandCoast + ClaimedSea) = ~67%

Very interesting! Suggestive of two distinct bugs - it is hard to imagine it being intentional given the Civilopedia's description. (Also suggestive of slightly un-modularized code in this area )

I'm tempted to put this to the test in my GOTM#5. I'm not at a point to decide yet, am currently at a very enjoyable greedy war-mongering stage But after that perhaps I should count the tiles then stay say 20 tiles short of that formula's limit. (Want to stay a bit short because it still seems likely that there is an occasional additional twist in the calculation, i.e. the thing which may somehow involve rivals in the threshold at times.)

 

[Aeson]

Remember that my ClaimedSea figure was a rough estimate. Also it might just be a formula that works in one situation (ratio of land/coastal to sea tiles) but not another. I think eyeballing it is still the most accurate form of predicting the domination limit. I wouldn't want a contrived formula of mine to mess up anybodys GOTM outcome. We still don't know if 2/3rds is the magic number the manual says it is. I guess what I'm trying to say is that I wouldn't use such a formula to determine my own GOTM domination limit without a lot more testing first.

 

[SirPleb]

Originally posted by Aeson
I guess what I'm trying to say is that I wouldn't use such a formula to determine my own GOTM domination limit without a lot more testing first.

Soon I will be at the stage where I should at least think about counting tiles. As that approaches I'm feeling more inclined to be cautious after all If I decide to play it out I think I'll do a tile count (tedious but not as bad on this map as previous months) and then use 2/3 of (land+coastal), using (land+coastal) for both numbers, leaving the sea part out.

 

[chiefpaco]

If you do, SirPleb, please don't do what I just did the other day while my brain was somewhere else. Don't count your land tiles. Count your sea & coasts & subtract that from your Land Area. Ooh, I wasted some time :crazyeyes

Oh, and BTW, just funny after re-reading my 1st post in this thread, I think we concluded that territory really has nothing to do with the domination calculation...

Still working with a larger map here...

 

[SirPleb]

Ouch, that hurts, counting them all. I figure on:
My count: Territory from Demographics, minus sea tiles.
Total count: Either take (totalmap - ocean - sea) or (mycount + unownedland + unownedcoastal), whichever looks faster.
Maybe I won't have to, not sure yet whether to shoot for an early win.

It sure doesn't seem that domination is related to territory, except as a short-cut for doing less counting

I'm tempted to try to write a utility to do the counting against a .sav file. Using a counting utility might be considered cheating for GOTM though. OTOH, it is almost sure to be more work to write a utility for the job vs. counting one map. (More fun though, and wins in time once re-used a bit )

 

[chiefpaco]

Ok, any more shortcuts I should be knowing? I just found an editor that lets you switch the player's civ. Using that you can find out the Land Area & territory scores of your opponents. I didn't know you could do that before today!

The editor I used was Civ3SavEdit - can we suggest utilities here? It's just for helping count land area, territory, and score.

Edit: Didn't want to have to post something new, but the C3MT is good too. I just took my map and exported it to a .bic. There's a map statistics button in his map editor that gives the land count. Now if only the water count was broken down to sea, coast, & ocean...

 

[chiefpaco]

Originally posted by David In Asia
will go over that thread now for my twopence worth......sorry to butt in on your discussion!

Hey. Opinions are welcome here too! I was just trying to save myself some typing!

Alright, here's my third test results from my only standard map, domination victory. Took a little while to get the tile counts and some could have a small error margin, but here goes:

First, the numbers:
Map Stats:
Total Land (From C3MT): 1318
Total Water (C3MT): 3682
Total Coast (Hand Count): 391
Total Land+Coast (summed): 1709

Year: 1782
My Land Area (F11 screen) (A): 1245
My Sea (Hand count): 121
My Land+Coast (A-Sea): 1124

Domination triggered by the beginning of the next turn, after I captured 1 city & 1 border expanded:
Year: 1784
My Land Area (A): 1262
My Sea : 121
My Land+Coast: 1141

My assumptions:
- Ocean, Sea, Territory Score, and opponents' Tiles do not factor into my civ's Domination calculation, so I ignored them, other than using them to derive other numbers.

Using the equation:
Domination % (D) = My (Land+Coast) / Total (Land+Coast)

in 1782
D = 1124/1709 = 65.8%

in 1784
D = 1141/1709 = 66.8%

These numbers support my Domination equation above. I'm still keeping a watch for the funny sea tile equation Aeson reported, but they weren't necessary in this case. Perhaps there are other factors we're not considering yet.

A further note:
I'm pretty sure I chose a 70% water coverage this map. Since the total land was 1318, that's something like 26% land. My point is, it is hard to tell even close to how many tiles you need each game for a domination victory with such a large change in the land mass.

An even further note:
Yes, I've done some checking to make sure the territory calc holds for bigger maps, later in the game & it still has worked.

 

[Beard Rinker]

I think there are some more conclusions that can be drawn from chiefpaco's 6 X 6 land test.

First I'll make a few unproven assumptions:
- Domination occurs when a civilization achieves 2/3 of something.
- All land tiles count towards the calculation.
- Unclaimed ocean tiles are not used in the calculation.

The question is what types of tiles are used in the calculation and when are they used. There are various possibilities:
- A tile type is always used.
- A tile type is only counted when it is claimed.
- A tile type is only counted when it is in a cities 21 tile boundary.

On the turn prior to domination the percentages where as follows:
1) Civ's land / Total Land = 72.2%
2) (Civ's Land + Civ's Coast) / Total Land and Coast = 65.6%
3) (Civ's Land + Civ's Coast) / Total Claimed Land and Coast = 71.2%
4) (Civ's Land + Civ's Cost + Civ's Ocean) / (Total Land and Coast + Claimed Ocean) = 67.2
5) (Civ's Land + Civ's Cost + Civ's Ocean) / (Total Claime Land, Coast and Ocean) = 72.6%

All of the above calculations except the land and coast calculation (2) where over the threshhold. After domination, the land and coast calculation crossed the threshhold.

This leads to the following conclustions:
- Land tiles alone are not used. There is something else. (Calculation 1)
- Ocean tiles are never used in the calculation. This includes not counting ocean tiles within a civilization's territory or 21 tile city boundary. (Calculation 4 & 5)
- Coastal tiles are always counted regardless of whether they are claimed or not. (Calculation 2 & 3)

It also supports chiefpaco's domination formula theory:
Domination% = My(land + coast) / Total(land + coast)

The remaining question is how are sea tiles counted. Anecdotal evidence seems to support that they are not counted at all.

 

[Beard Rinker]

On a side note, it is also entirely possible that coastal tiles are intentionally included in the domination calculation. After all, there is no 'Lake' tile, lakes appear as coastal tiles but behave slightly differently.

 

[SirPleb]

Nice work chiefpaco!!!

I don't know of any other utilities. But I haven't used them though so I have missed something. I did look around a bit last night, looking for anyone who has published discoveries about the save file format. I think that a couple of people have made good progress in decoding it, but they don't seem to be publishing that info So I don't think I'll write a tile counting program. I don't want to start from scratch working out the file format. I've now done my GOTM5 count the hard way. (Actually, I counted a representative 1/3 of the map and multiplied. It sure is painful and error prone counting tiles!)

Beard Rinker: That's sure giving them the benefit of the doubt! I wondered about that (it being intentional) a while back. But I couldn't reconcile the idea of galleys sailing the coastal tiles with them being part of "the world's land surface". Of course, it could be that the Civilopedia has the bug, not the program. That's my preferred approach when possible - if the docs and program don't match, fix the docs!

 

[Duke of Marlbrough]

Nice to see most of my main HOF contributors are taking big parts in exploring the limits of the game.

 

[SirPleb]

I just went back to check my GOTM#5, to see how close I was to the Domination threshold.

The results are not exact but are strongly consistent with the 2/3 of (land+coastal) theory.

Before stopping my expansion I decided to do some counting. I intended to count all sea+ocean tiles on the map, to determine its land+coastal. I didn't finish the count My eyes got too blurry and I hated the task. But I counted about 20% and felt it was enough to make a very good guess - distribution on this map is uniform enough to treat a part of it as representative.

(BTW, I'm not putting actual numbers in this note because that could be GOTM spoiler info.)

I then took 2/3 of that number minus 60 as my limit. (Subtracted 60 to allow for error in the counting and in the guesstimate.)

After that, I periodically counted all my owned sea tiles, subtracted them from the territory number on the Demographics screen, and checked how I was doing and eventually stopped growing at just a hair over my target number.

In my testing just now, it turns out I was 81 tiles (+/- 2) from hitting Domination.

That's only 20 off my guesstimate, easily within the margin of error for my incomplete count. But 20 is substantially smaller than the error would be if sea tiles were included in either part of the calculation. (The sea tiles within my sphere of influence number in the hundreds.)

So there's another strong indication that it is:
2/3 of total(land+coastal) within one's sphere of influence